Probability Density Function Distance-Based Augmented CycleGAN for Image Domain Translation with Asymmetric Sample Size

Many image-to-image translation tasks face an inherent problem of asymmetry in the domains, meaning that one of the domains is scarce—i.e., it contains significantly less available training data in comparison to the other domain. There are only a few methods proposed in the literature that tackle th...

Full description

Saved in:
Bibliographic Details
Main Authors: Lidija Krstanović, Branislav Popović, Sebastian Baloš, Milan Narandžić, Branko Brkljač
Format: Article
Language:English
Published: MDPI AG 2025-04-01
Series:Mathematics
Subjects:
Online Access:https://www.mdpi.com/2227-7390/13/9/1406
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850279218569543680
author Lidija Krstanović
Branislav Popović
Sebastian Baloš
Milan Narandžić
Branko Brkljač
author_facet Lidija Krstanović
Branislav Popović
Sebastian Baloš
Milan Narandžić
Branko Brkljač
author_sort Lidija Krstanović
collection DOAJ
description Many image-to-image translation tasks face an inherent problem of asymmetry in the domains, meaning that one of the domains is scarce—i.e., it contains significantly less available training data in comparison to the other domain. There are only a few methods proposed in the literature that tackle the problem of training a CycleGAN in such an environment. In this paper, we propose a novel method that utilizes pdf (probability density function) distance-based augmentation of the discriminator network corresponding to the scarce domain. Namely, the method involves adding examples translated from the non-scarce domain into the pool of the discriminator corresponding to the scarce domain, but only those examples for which the assumed Gaussian pdf in VGG19 net feature space is sufficiently close to the GMM pdf that represents the relevant initial pool in the same feature space. In experiments on several datasets, the proposed method showed significantly improved characteristics in comparison with a standard unsupervised CycleGAN, as well as with Bootstraped SSL CycleGAN, where translated examples are added to the pool of the discriminator corresponding to the scarce domain, without any discrimination. Moreover, in the considered scarce scenarios, it also shows competitive results in comparison to fully supervised image-to-image translation based on the pix2pix method.
format Article
id doaj-art-3b2af63afb42493c910e1d052a4035a3
institution OA Journals
issn 2227-7390
language English
publishDate 2025-04-01
publisher MDPI AG
record_format Article
series Mathematics
spelling doaj-art-3b2af63afb42493c910e1d052a4035a32025-08-20T01:49:11ZengMDPI AGMathematics2227-73902025-04-01139140610.3390/math13091406Probability Density Function Distance-Based Augmented CycleGAN for Image Domain Translation with Asymmetric Sample SizeLidija Krstanović0Branislav Popović1Sebastian Baloš2Milan Narandžić3Branko Brkljač4Department of Fundamental Disciplines in Engineering, Faculty of Technical Sciences, University of Novi Sad, Trg Dositeja Obradovića 6, 21000 Novi Sad, SerbiaDepartment of Power, Electronic and Telecommunication Engineering, Faculty of Technical Sciences, University of Novi Sad, Trg Dositeja Obradovića 6, 21000 Novi Sad, SerbiaDepartment of Production Engineering, Faculty of Technical Sciences, University of Novi Sad, Trg Dositeja Obradovića 6, 21000 Novi Sad, SerbiaDepartment of Power, Electronic and Telecommunication Engineering, Faculty of Technical Sciences, University of Novi Sad, Trg Dositeja Obradovića 6, 21000 Novi Sad, SerbiaDepartment of Power, Electronic and Telecommunication Engineering, Faculty of Technical Sciences, University of Novi Sad, Trg Dositeja Obradovića 6, 21000 Novi Sad, SerbiaMany image-to-image translation tasks face an inherent problem of asymmetry in the domains, meaning that one of the domains is scarce—i.e., it contains significantly less available training data in comparison to the other domain. There are only a few methods proposed in the literature that tackle the problem of training a CycleGAN in such an environment. In this paper, we propose a novel method that utilizes pdf (probability density function) distance-based augmentation of the discriminator network corresponding to the scarce domain. Namely, the method involves adding examples translated from the non-scarce domain into the pool of the discriminator corresponding to the scarce domain, but only those examples for which the assumed Gaussian pdf in VGG19 net feature space is sufficiently close to the GMM pdf that represents the relevant initial pool in the same feature space. In experiments on several datasets, the proposed method showed significantly improved characteristics in comparison with a standard unsupervised CycleGAN, as well as with Bootstraped SSL CycleGAN, where translated examples are added to the pool of the discriminator corresponding to the scarce domain, without any discrimination. Moreover, in the considered scarce scenarios, it also shows competitive results in comparison to fully supervised image-to-image translation based on the pix2pix method.https://www.mdpi.com/2227-7390/13/9/1406CycleGANdomain translationselective data augmentationbootstrappingpdf distance
spellingShingle Lidija Krstanović
Branislav Popović
Sebastian Baloš
Milan Narandžić
Branko Brkljač
Probability Density Function Distance-Based Augmented CycleGAN for Image Domain Translation with Asymmetric Sample Size
Mathematics
CycleGAN
domain translation
selective data augmentation
bootstrapping
pdf distance
title Probability Density Function Distance-Based Augmented CycleGAN for Image Domain Translation with Asymmetric Sample Size
title_full Probability Density Function Distance-Based Augmented CycleGAN for Image Domain Translation with Asymmetric Sample Size
title_fullStr Probability Density Function Distance-Based Augmented CycleGAN for Image Domain Translation with Asymmetric Sample Size
title_full_unstemmed Probability Density Function Distance-Based Augmented CycleGAN for Image Domain Translation with Asymmetric Sample Size
title_short Probability Density Function Distance-Based Augmented CycleGAN for Image Domain Translation with Asymmetric Sample Size
title_sort probability density function distance based augmented cyclegan for image domain translation with asymmetric sample size
topic CycleGAN
domain translation
selective data augmentation
bootstrapping
pdf distance
url https://www.mdpi.com/2227-7390/13/9/1406
work_keys_str_mv AT lidijakrstanovic probabilitydensityfunctiondistancebasedaugmentedcycleganforimagedomaintranslationwithasymmetricsamplesize
AT branislavpopovic probabilitydensityfunctiondistancebasedaugmentedcycleganforimagedomaintranslationwithasymmetricsamplesize
AT sebastianbalos probabilitydensityfunctiondistancebasedaugmentedcycleganforimagedomaintranslationwithasymmetricsamplesize
AT milannarandzic probabilitydensityfunctiondistancebasedaugmentedcycleganforimagedomaintranslationwithasymmetricsamplesize
AT brankobrkljac probabilitydensityfunctiondistancebasedaugmentedcycleganforimagedomaintranslationwithasymmetricsamplesize